Using Quantitative EEG to Manage Blood Pressure in a Child With Carotid ECMO Cannulation: Case Report and Proof of Concept

Abstract

Pediatric extracorporeal membrane oxygenation (ECMO) carries a high risk of arterial ischemic stroke (AIS). AIS in patients with right common carotid artery (RCCA) cannulation may be driven by inadequate cross-lateral perfusion from the left hemisphere during hypotensive episodes. We present the case of a 2-year-old on ECMO with RCCA cannulation who had electroencephalographic (EEG) evidence of right hemispheric ischemia yet survived ECMO without cortical infarct and minimal neurologic deficits. Neuroimaging after decannulation showed right carotid narrowing that resolved after 2 weeks. Analysis of time-synchronized quantitative EEG metrics and invasive blood pressure monitoring showed a weak but present correlation (r2 = 0.18), indicating covariant fluctuation in blood pressure and EEG asymmetry. This report affirms EEG as a tool to identify insufficient cross-lateral cerebral perfusion with RCCA cannulation as a potential mechanism of AIS pathogenesis in ECMO. Further research is needed to optimize EEG-guided hemodynamic management in patients with RCCA cannulation.

Arterial ischemic stroke (AIS) is a devastating complication of pediatric extracorporeal membrane oxygenation (ECMO). ¹ Understanding the mechanisms of AIS in ECMO is critical to improving neurologic outcomes.^2–5 Right common carotid artery (RCCA) cannulation is associated with right hemispheric AIS. ² The propensity for AIS ipsilateral to RCCA cannulation is not fully explained by a thromboembolic mechanism. ³

Early AIS recognition during ECMO is challenging as the need for sedation limits the neurological exam. ⁴ Quantitative EEG assists in detecting background asymmetries associated with AIS.^2,5,6 Background EEG asymmetries can be quantified and visualized using the Correlate Of Injury to the Nervous System (COIN) index. ⁷

We report a case of pediatric ECMO with RCCA cannulation and EEG abnormalities suggestive of right hemispheric ischemia with no cortical infarct identified on serial imaging. We retrospectively analyzed EEG data using COIN to measure the degree of focal neurophysiologic dysfunction and time-synchronized it with continuous blood pressure values to see whether COIN fluctuations co-occurred with hemodynamic fluctuations.

Methods

Clinical Data

Electronic health record was reviewed for hospital course, physical examination findings, laboratory data, neuroimaging, and EEG reports. Institutional review board approval was obtained and parents provided written consent for publication of clinical findings. Reporting of information is limited to generalities to minimize identification risk.

Quantitative EEG and Statistical Analysis

EEG files were processed using Persyst (Persyst, Prescott, AZ) for left and right hemispheric Fast Fourier Transform spectrogram, Asymmetry-Relative Spectrogram, and amplitude integrated EEG. EEG data were also downloaded from Natus (Natus Medical Inc, Middleton, WI), deidentified, and analyzed in MATLAB (MathWorks, Natick, MA) to obtain COIN values using our previously reported algorithm. ⁷ A COIN value of 0 implies a symmetric EEG, and negative values are observed with power attenuation in the 6- to 16-Hz range across 1 or more channels.

Continuous blood pressure data were downloaded from BedMaster (Excel Medical, Jupiter, FL) and time-synchronized with EEG data on MATLAB. To assess whether fluctuations in mean arterial pressure (MAP) correlated with fluctuations in COIN, we subtracted a 3-hour moving average from a 1-hour moving average of both MAP and COIN trends. Data were visualized using scatterplots, and a Pearson correlation coefficient was calculated to assess the strength of correlation.

Case Description

A 2-year-old male child with no medical history presented with vomiting and melena. Leukocytosis, thrombocytopenia, and acute renal failure led to a diagnosis of Shiga toxin–associated hemolytic uremic syndrome. At admission, he had no focal examination findings and was following commands. He started renal replacement on day 6 of illness and was intubated on day 9 of illness for pulmonary hemorrhage. He developed biventricular myocardial failure requiring inotropic infusions.

On day 12 of illness, he suddenly developed pulseless electrical arrest and was emergently cannulated onto ECMO through the RCCA and internal jugular vein. The distal RCCA was not ligated. Neurologic examination on continuous neuromuscular blockade following cannulation showed equally reactive pupils. Serum lactate peaked at 11.9 mmol/L and improved to 1.8 mmol/L within 12 hours of cannulation; eculizumab was given to treat the underlying disease and myocardial failure.

Over the first 24 hours of ECMO, he developed left atrial hypertension requiring cardiac catheterization for balloon atrial septostomy. Following catheterization, EEG showed a background asymmetry, with the right hemisphere showing relative attenuation in alpha and beta power (8-20 Hz) and augmentation in delta power (1-4 Hz). Pupils remained equally reactive. He underwent a beside head computed tomography on ECMO day 2 that was normal (Figure 1A). Because of a concern for right hemispheric ischemia, ECMO circuit flows and vasoactive infusions were modulated to maintain a higher MAP goal of >75 mm Hg, which was achieved for the remainder of the ECMO run. There were no appreciable changes in raw EEG output relative to blood pressure fluctuations.

Figure 1.

(A) Noncontrast head CT during ECMO demonstrates no acute intracranial abnormality. (B-D) MR brain without contrast 1 day after ECMO de-cannulation. (B) Axial and (C) coronal DWI demonstrates focus of reduced diffusion in the left lateral pons (arrows), suggestive of infarction. Active demyelination, which occurs in that area, was thought less likely given the clinical context. (D) Time-of-flight MRA neck performed at the same time reveal irregularity of the lower right common carotid artery (long arrow) at the presumed site of ECMO cannulation and arterial ligation. There is asymmetrically decreased caliber of the right common and internal carotid arteries (arrowheads) compared to the left. (E) Time-of-flight MRA neck performed 16 days after decannulation demonstrates resolution of the right carotid narrowing. Both common and internal carotid arteries are now equal in caliber. The area of carotid cannulation no longer appears as irregular, although the region is not well-evaluated as it is on the edge of the field of view. CT, computed tomography; DWI, diffusion-weighted imaging; ECMO, extracorporeal membrane oxygenation; MRA, magnetic resonance angiography.

Four days after ECMO initiation, myocardial function improved, and the patient was uneventfully decannulated with RCCA reconstruction. Examination while sedated and off paralytic after decannulation was notable for no verbal output, no command following, intermittent dysconjugate gaze with left eye abduction, spontaneous movement on his right side only, and increased tone on the left compared with the right.

Magnetic resonance imaging of the brain with time-of-flight angiography of the cervical and intracranial vasculature was obtained 1 day after decannulation, demonstrating a small sub-acute left pontine infarct and diffuse right carotid arterial narrowing with a focal irregularity at the presumed site of cannulation (Figure 1B-D). Repeat vascular imaging 2 weeks later showed resolution of the right carotid narrowing (Figure 1E). Discharge examination showed mild left elbow contracture and was otherwise normal with full symmetric strength and age-appropriate mental status.

COIN and Blood Pressure Analysis

Post hoc COIN analysis on EEG following cardiac catheterization showed a persistent attenuation on the right with COIN values that remained between 0 and −20 for the duration of recording, suggesting against the presence of a large-volume AIS. Persyst readout showed relative attenuation of alpha power and augmentation of delta power in the right hemisphere, with overall amplitude higher on the right than left, that qualitatively appears to cofluctuate with blood pressure (Figure 2). Correlation coefficient (r²) of raw COIN against MAP was 0.001, and of COIN fluctuations against MAP fluctuations was 0.18 (Supplemental Figure 1).

Figure 2.

(A) Raw EEG tracing from 4 representative 4-second epochs in the period before and after noncontrast head CT. (B) COIN visualization from representative EEG epochs shows persistent right hemispheric attenuation across multiple bipolar channels. (C) Continuous COIN trend shows dynamic trend in right and left hemispheric COIN values over 36 hours of EEG recording, with summary COIN values maintained above −20 for the majority of the recording. Recording time is relative to ECMO start, with pause in recording during head CT. (D) Continuous blood pressure shows MAP maintained above 65 mm Hg during the pre-CT period, and above 75 mm Hg in the post-CT period. (E) Persyst trends derived from the EEG tracing show relative attenuation of alpha and beta power and augmentation of delta power in the right hemisphere. aEEG, amplitude-integrated electroencephalography; A-RS, asymmetry–relative spectrogram; COIN, Correlate Of Injury to the Nervous system; CT, computed tomography; ECMO, extracorporeal membrane oxygenation; EEG, electroencephalography; FFT, fast Fourier transform; L, left; PS, parasagittal; R, right; T, temporal.

Discussion

This case provides an avenue to study neurophysiology-informed hemodynamic management in pediatric ECMO with RCCA cannulation. ² The EEG features during ECMO and the examination findings after decannulation clinically suggested large vascular territorial ischemia of the right hemisphere. However, brain imaging obtained both during and after the ECMO run were negative for right hemispheric AIS. We believe that right carotid narrowing resulted in inadequate right hemispheric perfusion, and that blood pressure augmentation may have sufficiently supported collateral perfusion to prevent infarction.

The finding of transient diffuse right internal carotid artery narrowing implies vascular reactivity possibly related to RCCA cannulation. We do not know whether the narrowing occurred during cannulation and persisted during ECMO or whether it was a result of manipulation of the RCCA during decannulation. However, similar findings have been described previously,^8,9 and a transient vasculopathy causing perfusion deficits may partially explain the proclivity for AIS ipsilateral to carotid cannulation. ²

AIS remains a serious and feared complication of pediatric ECMO in children. The rate of ECMO utilization is increasing as its use-indication expands, yet AIS rates remain high. Carotid occlusion is a described risk factor for ipsilateral AIS. ² Because of the potential for ipsilateral hypoperfusion with carotid cannulation, AIS genesis and severity may be mitigated by clinical interventions to augment blood pressure and prevent sustained relative hypotensive episodes.

COIN may serve as a helpful adjunct in neurophysiologic monitoring for AIS detection during ECMO. Previous evaluation of COIN has shown that a threshold of −18 has the best accuracy for detection of large volume AIS.^7,10 In this case, COIN mostly stayed between −5 and −15. A COIN value in this range may indicate “at risk” brain tissue for which neurologic evaluation is warranted. ¹¹ The COIN visualization also provides an appraisal of focal background changes to single channels, which can be challenging to appreciate on raw EEG and localize using aggregated quantitative EEG displays.

This case highlights how multimodal monitoring with EEG and blood pressure synchronization provides more nuanced clinical information that informs clinical management compared with overreliance on a single modality. COIN may enable quantitative study of the potential hemodynamic effect on ischemic neurophysiology. Future work will compare the effect of severe hypotension on COIN in patients with and without carotid cannulation.

A better understanding of EEG markers of AIS pathogenesis during ECMO is needed to mitigate AIS disability and mortality. This case report highlights potential avenues to study AIS pathogenesis with carotid cannulation and affirms the role of quantitative EEG monitoring for AIS prevention during ECMO.

Supplemental Material

sj-tiff-1-jcn-10.1177_08830738261451031 - Supplemental material for Using Quantitative EEG to Manage Blood Pressure in a Child With Carotid ECMO Cannulation: Case Report and Proof of Concept

Supplemental material, sj-tiff-1-jcn-10.1177_08830738261451031 for Using Quantitative EEG to Manage Blood Pressure in a Child With Carotid ECMO Cannulation: Case Report and Proof of Concept by Natalie Pierson, Natalia Lopez, Yi Li, Amar Nijagal, Kathryn Finlay, Loren D. Sacks, Adam L. Numis, Christine K. Fox and Mauro Caffarelli in Journal of Child Neurology

Supplemental Material

sj-docx-2-jcn-10.1177_08830738261451031 - Supplemental material for Using Quantitative EEG to Manage Blood Pressure in a Child With Carotid ECMO Cannulation: Case Report and Proof of Concept

Supplemental material, sj-docx-2-jcn-10.1177_08830738261451031 for Using Quantitative EEG to Manage Blood Pressure in a Child With Carotid ECMO Cannulation: Case Report and Proof of Concept by Natalie Pierson, Natalia Lopez, Yi Li, Amar Nijagal, Kathryn Finlay, Loren D. Sacks, Adam L. Numis, Christine K. Fox and Mauro Caffarelli in Journal of Child Neurology